You use go_action&action=ocrexclude
to exclude the current page from any text recognition.
You should use this option if certain pages within files only contain images since the OCR text recognition does not provide usable text and does not work quickly in the case of images.